You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The AI-credits cluster is expanding (Day 4, grew from 4 → 6+ workflows) per the Agent Performance Report (#38626) and Daily Workflow Audit (#38444). Heavy aggregators reliably brush the ~1000 daily AIC cap and abort with CAPIError 429 Maximum AI credits exceeded (1003/1000) after retries (e.g. Daily Ambient Context Optimizer, Package Specification Extractor #38576 at 1,023.9 AIC). The audit recommendation rec-ai-credits-soft-cap-guard is still not shipped, and #38624 only raises the cap for one workflow rather than addressing the root cause.
Scope
Before the heaviest aggregation turns, check remaining AI credits and degrade gracefully to partial output + noop/report_incomplete instead of hard-aborting with a 429. Start with the repeat offenders: Daily Ambient Context Optimizer and Package Specification Extractor.
Expected Impact
Converts hard 429 failures (red runs + [aw] ... failed issue noise) into clean partial completions.
Description
The AI-credits cluster is expanding (Day 4, grew from 4 → 6+ workflows) per the Agent Performance Report (#38626) and Daily Workflow Audit (#38444). Heavy aggregators reliably brush the ~1000 daily AIC cap and abort with
CAPIError 429 Maximum AI credits exceeded (1003/1000)after retries (e.g. Daily Ambient Context Optimizer, Package Specification Extractor #38576 at 1,023.9 AIC). The audit recommendationrec-ai-credits-soft-cap-guardis still not shipped, and #38624 only raises the cap for one workflow rather than addressing the root cause.Scope
Before the heaviest aggregation turns, check remaining AI credits and degrade gracefully to partial output +
noop/report_incompleteinstead of hard-aborting with a 429. Start with the repeat offenders: Daily Ambient Context Optimizer and Package Specification Extractor.Expected Impact
[aw] ... failedissue noise) into clean partial completions.Suggested Agent
New Agent or copilot-swe-agent (engine/driver-level change).
Estimated Effort
Medium (1–4 hours).
Data Source
DeepReport run 2026-06-11; Daily Workflow Audit #38444 (
rec-ai-credits-soft-cap-guard), Agent Performance Report #38626.